Par4Sem

Par4Sem can be installed in three different ways. While the first way allows to run only the frontend component from any web server (or even locally using a browser), the other two approach requires installation of Docker.

1. Minimal installation:
   • Download the frontend zip file from here
   • Unzip the file and either 1) put the whole file in your web-server (such as apache) or 2) open the index.html file in your browser.
   • If you deploy the file in a webserver, access it with the context name (if you deploy the file under the folder par4Sem, access it as (http://WEBADDRESS/par4sem)
   • Voila you are done 😄      
2. Complete local installation Use resources and database dumps on local Machine. Follow this approach if you like to install the complete Par4Sem software in your local server. Requirements: To run the whole application you need to fulfill the following:
   • Docker Engine and docker compose
   • RAM ~4GB
   • Hard Disk: ~100GB (for different resources and database dumps)

   All the resource files such as phrase2vec embeddings, OpenNLP models for parts-of-speech tagging, tokenizations, and sentence splitting, Wikipedia frequency counts are available here. Database dumps for the different paraphrase resources such as PPDB 2.0, JoBimText DT, WordNet, SimplePPDB and CWI datasets are available here. here. To start Par4Sem:

   • Download the docker-compose file.
   • Download the database dump, unzip and put them under dumps folder
   • download the resources file, unzip and put them under data folder
   • start the docker as docker-compose up -d
3. From source code
   • Clone the source code from GutHub: git clone https://github.com/uhh-lt/par4sem.git
   • run mvn clean install inside par4sem directory
   • copy the DockerFile, and the web/targte/par4sem.war file to a directory of your choice
   • Run docker build -t par4sem .
   • Follow Step two above but make sure to edit the docker-compose file to point to the new image you build.